Skip to main content

Full text of "Chandra Detection of X-ray Absorption from Local Warm/Hot Gas"

See other formats


T. Fang 1 ' 2 , C. Canizares 2 , K. Sembach 3 , H. Marshall 2 , J. Lee 2 , D. Davis 4 

1. Dept. of Physics, Carnegie Mellon Univ., 5000 Forbes Ave., Pittsburgh, PA 15213 

2. MIT, Center for Space Research, 70 Vassar St., Cambridge, MA 02139 

3. STScI, 3700 San Martin Drive, Baltimore, MD 21218 

4. GSFC, Code 661., GLAST SSC, Greenbelt, MD 20771 

Abstract Recently, with the Chandra X-ray Telescope we have detected several 
local X-ray absorption lines along lines-of-sight towards distant quasars. 
These absorption lines are produced by warm/hot gas located in local 
intergalactic space and/or in our Galaxy. I will present our observations 
and discuss the origin of the X-ray absorption and its implications in 
probing the warm/hot component of local baryons. 

1. Introduction 

The cosmic baryon budget at low and high redshift indicates that a 
large fraction of baryons in the local universe have so far escaped de- 
tection (e.g., Fukugita, Hogan, & Peebles 1998). While there is clear 
evidence that a significant fraction of these "missing baryons" (between 
20-40% of total baryons) lie in photoionized, low-redshift Lya clouds 
(Penton, Shull, h Stocke 2000), the remainder could be located in in- 
tergalactic space with temperatures of 10 5 — 10 7 K (warm-hot inter- 
galactic medium, or WHIM). Resonant absorption from highly- ionized 
ions located in the WHIM gas has been predicted based on both an- 
alytic studies of structure formation and evolution (Shapiro & Bah- 
call 1981;Perna & Loeb 1998;Fang & Canizares 2000) and cosmic hy- 
drodynamic simulations (Hellsten, Gnedin, & Miralda-Escude 1998; Cen 
& Ostriker 1999a;Dave et al. 2001;Fang, Bryan, & Canizares 2002). Re- 
cent discovery of O VI absorption lines by the Hubble Space Telescope 
(HST) and the Far Ultraviolet Spectroscopic Explorer (FUSE) (see, e.g., 
Tripp &; Savage 2000) indicates that there may be a significant reser- 
voir of baryons in O VI absorbers. While Li-like O VI probes about 
~ 30 - 40% of the WHIM gas (Cen et al. 2001;Fang & Bryan 2001), the 



remaining ~ 60 — 70% is hotter and can only be probed by ions with 
higher ionization potentials, such as H- and He-like Oxygen, through 
X-ray observation. 

Recently, with Chandra Low Energy Transmission Grating Spectrom- 
eter (LETGS) we detected resonance absorption lines from H- and He- 
like Oxygen in the X-ray spectra of background quasars, namely PKS 2155- 
304 and 3C 273. The detected lines can be categorized into (1) those at 
z ~ and (2) one redshifted intervening system. In this paper, we will 
discuss these detections and their implications for the physical properties 
of the hot gases that give rise to these absorption features. 

Table 1: Fitting parameters of the X-ray absorption Lines 

PKS 2155-304 3C 273 

viii Lya 

O vil Hea 

O vn Hea 

cz(km s _1 ) 

20 02 +u ulb 
zu.uz_ 015 

16624 ± 237 

2i.6itr i 



Line Width a 

< 0.039 

< 0.027 

< 0.020 

Line Flux b 
EW (mA) 



4 2+ 1 ' 8 


28.4lJ 2 2 5 





a. 90% upper limit of the line width <r, in units of A. 

b. Absorbed line flux in units of 10 -5 photons cm _2 s _1 . 

2. Data Reduction 

PKS 2155-304 and 3C 273 are bright extragalactic X-ray sources used 
as Chandra calibration targets. They were observed with the Chandra 
LETG-ACIS (the observations ids for PKS 2155 are 1703, 2335, 3168; 
and the ids for 3C 273 are 1198, 2464, 2471). For detailed data analysis, 
we refer to Fang et al. 2002. We found all continua are well described 
by a single power law absorbed by Galactic neutral hydrogen. 

After a blind search for any statistically significant absorption fea- 
tures, several absorption features with S/N > 4 were detected in the 
spectra of both quasars in the 2-42 A region of the LETGS spectral 
bandpass (Figure 1). These features were subsequently fit in ISIS (Houck 
& Denicola 2000). 

3. Discussion 

3.1 PKS 2155-304 

The absorption feature at ~ 21.6 A was reported by Nicastro et 
al. (2002) in the LETGS-HRC archival data. We concentrate on the 

Chandra Detection of X-ray Absorption from Local Warm/Hot Gas 


absorption feature which appears at 20.02 A (619 eV). Considering 
cosmic abundances and oscillator strengths for different ions, O VIII 
Lya is the only strong candidate line between 18 and 20 A , the mea- 
sured wavelength de-redshifted to the source. It is plausible that the 
20 A absorption is due to O VIII Lya in a known intervening sys- 
tem at cz « 16,734 km s -1 . With HST, Shull et al. (1998) discov- 
ered a cluster of low metallicity H I Lya clouds along the line-of-sight 
(LOS) towards PKS 2155-304, most of which have redshifts between 
cz = 16, 100 km s" 1 and 18, 500 km s _1 . Using 21 cm images from the 
Very Large Array ( VLA) , they detected a small group of four H I galax- 
ies offset by ~ 400 — 800 hj Q kpc from the LOS, and suggested that the 
H I Lya clouds could arise from gas associated with the group (We use 
Hq = 701i7o km s Mpc throughout the paper). 

19.0 19.5 20.0 20.5 21.0 21.5 22.0 22.5 30.0 20.5 21.0 21.5 22.0 22.5 23.0 

Wavelength (A) Wavelength (A) 

Figure 1 The Chandra LETG-ACIS spectra of (a) PKS 2155 and (b) 
3C273. The dashed lines are the fitted spectra. The average la error 
bar plotted on the right-bottom of each panel is based on statistics only. 

Taking the absorption line to be O VIII Lya, we estimate the column 
density is N(0 VIII) ~ 9.5 x 10 15 cm" 2 if the line is unsaturated. We 
can constrain the density of the absorbing gas, assuming it is associated 
with the intervening galaxy group. Since the line is unresolved, a lower 
limit of n b > (1.0 x 10~ 5 cm" 3 ) 1 can be obtained. Here Zo.i 

is the metallicity in units of 0.1 solar abundance, /0.5 is the ionization 
fraction in units of 0.5, and lg is the path length in units of 8h^ Mpc. A 
more reasonable estimate of the path length comes from the mean pro- 
jected separation of ~1 Mpc for the galaxies in the group, which gives 
n& ~ 7.5 x 10~ 5 cm" 3 ZqI/q^. This implies a range of baryon over- 
density (5b ~ 50 - 350) over the cosmic mean (rib) = 2.14 x 10~ 7 cm" 3 . 
Interestingly, Shull et al.(1998) estimate an overdensity for the galaxy 
group of 5 ga i ~100. 


In the case of pure collisional ionization, temperature is the only pa- 
rameter of importance over a wide range of density so long as the gas is 
optically thin. The O VIII ionization fraction peaks at 0.5, and exceeds 
0.1 for temperatures T ~ 2 - 5 x 10 6 K. Using CLOUDY (Ferland et 
al. 1998) we find that photoionzation by the cosmic UV/X-ray back- 
ground is not important for > 10~ 5 cm -3 . CLOUDY calculations of 
the column density ratios between other ions and O VIII also show that 
T £ 10 6 - 4 K. 

Assuming conservative upper limits of Z ^ 0.5Z Q and an O VIII ion- 
ization fraction of / ^ 0.5 , and a path length of Az <^ 0.116, we estimate 
n h (0 VIII) £ O.OOShfQ 1 . This is about 10% of the total baryon fraction, 
or about 30-40% of the WHIM gas, if the WHIM gas contains about 
30-40% of total baryonic matter. This baryon fraction is consistent with 
the prediction from Perna & Loeb (1998) based on a simple analytic 

3.2 3C 273 

Based on the detected line equivalent width (W\) and non-detection 
of local O VII He/3 line at 18.6288^4, we estimate the column density 
N(0 VII) = 1.8±o;7 x 10 16 cm" 2 if the line is unsaturated 1 . The O VII 
indicates the existence of the gas with high temperatures ~ 10 6 K. The 
extremely high temperatures imply that O VII is unlikely to to be pro- 
duced in the nearby interstellar medium (ISM) , although we cannot rule 
out the possibility of an origin from supernova remnants. It is more 
plausible that the He-like Oxygen is produced in distant, hot halo gas, 
or even in the Local Group (LG). 

3.2.1 Local Group Origin? 
We can constrain the density of the absorbing gas, assuming it is as- 
sociated with the Local Group. The path length can be set equal to 
the distance to the boundary of the Local Group, where the gas be- 
gins to participate in the Hubble flow. Assuming a simple geometry for 
the Local Group, the path length is ~ 1 Mpc (see the following text). 
Adopting the 90% lower limit of the O VII column density, this gives 
rib > (2-2 x 10 -5 cm~ 3 ) Zq I/i 1 ^ 1 , where Z0.2 is the metallicity in 
units of 0.2 solar abundance, /1 is the ionization fraction, and l\ is the 
path length in units of 1 Mpc. 

Assuming spherical symmetry and isothermality, a model character- 
izing the distribution of the Local Group gas is given by the standard 

1 Rasmussen et al.(2002) also reported the detection of a similar feature with XMM-Newton 
in this conference. 

Chandra Detection of X-ray Absorption from Local Warm/Hot Gas 


/?-model. We adopt a simplified geometry model of the LG, where the 
LG barycenter is located along the line connecting M31 and the Milky 
Way, at about 450 kpc away from our Galaxy (Rasmussen & Peder- 
sen 2001). At about r ~ 1200 kpc the gravitational contraction of the 
Local Group starts to dominate the Hubble flow, and this was defined 
as the boundary of the Local Group (Courteau & van den Bergh 1999). 
Based on this simple model we estimate the column density of O VII by 
integrating the O VII number density no vil along this path length. We 
find a tight upper limit of Local Group temperature T < 1.2 x 10 6 K; at 
temperatures higher than 1.2 x 10 6 K, the O VII ionization fraction drops 
quickly and O VIII starts to dominate. We also find that the tempera- 
ture of the Local Group should be higher than 2.3 x 10 5 K. To satisfy the 
observed O VII column density we find that the gas distribution should 
have a rather flat core with r c > 100 kpc. 

3.2.2 Hot Halo Gas? 
Strong O VI absorption (logN(0 VI) = 14.73 ± 0.04) along the sight 
line towards 3C 273 was detected with FUSE between -100 and +100 
km s _1 (Sembach et al. 2001). This absorption probably occurs in the 
interstellar medium of the Milky Way disk and halo. Absorption fea- 
tures of lower ionization species are also present at these velocities. O VI 
absorption is also detected between +100 and +240 km s _1 in the form 
of a broad, shallow absorption wing extending redward of the primary 
Galactic absorption feature with logN(0 VI) = 13.71. The O VI ab- 
sorption wing has been attributed to hot gas flowing out of the Galactic 
disk as part of a "Galactic chimney" or "fountain" in the Loop IV and 
North Polar Spur regions of the sky. Alternatively, the wing might be 
remnant tidal debris from interactions of the Milky Way and smaller 
Local Group galaxies (Sembach et al. 2001). It is possible to associate 
the Chandra-detected O VII absorption with these highly ionized metals 
detected by FUSE. Here we discuss several scenarios: 

(1). The O VII is related to the primary O VI feature: In this case, 
logN(0 VI) = 14.73, and log[N(0 VII)/N(0 VI)] ~ 1.5, assuming 
N(0 VII) = 1.8 x 10 16 cmr 2 . This is within about a factor of 2 of 
the O vn/0 vi ratio observed for the PKS 2155-304 absorber and is 
consistent with the idea that the gas is radiatively cooling from a high 
temperature (Heckman et al. 2002). This possibility is appealing since 
the centroids of the O vi and O vn absorption features are similar 
(~ 6 + 10 km s _1 versus —26 ± 140 km s™ 1 ), and the width of the 
resolved O vi line (FWHM ~ 100 km s _1 ) is consistent with a broad 
O vn feature. However, this possibility also has drawbacks that the 
predicted O vm column density is too high and the amount of C iv 
predicted is too low. 


(2) . The O VII is related to the O VI wing: In this case, the O vn 
is associated only with the O vi absorption "wing". This seems like 
a reasonable possibility. Then log[N(0 VII) /N(0 VI)] ~ 2.5. In colli- 
sions! ionization equilibrium, this would imply a temperature of > 10 6 
K (Sutherland & Dopita 1993). The non-detection of O VIII Lya ab- 
sorption requires the temperatures lower than ~ 10 6 ' 3 K. 

(3) . The O VII is related to none of the O VI: In this case, the tem- 
perature should be high enough to prevent the production of O VI ions. 
We can reach the similar conclusions to those in situation (2). 

We thank members of the MIT/CXC team for their support. This 
work is supported in part by contracts NAS 8-38249 and SAO SV1- 
61010. KRS acknowledges financial support through NASA contract 
NAS 5-32985 and Long Term Space Astrophysics grant NAG5-3485. 


Cen, R. & Ostriker, J.P. 1999a, ApJ, 514, 1 
Cen, R. & Ostriker, J. P. 1999b, ApJ, 519, L109 

Cen, R., Tripp, T.M., Ostriker, J.P. & Jenkins, E.B. 2001, ApJ, 559, L5 

Courtcau, S. & van den Bergh, S. 1999, AJ, 118, 337 

Dave, R. et al. 2001, ApJ, 552, 473 

Fang, T. & Bryan, G. L. 2001, ApJ, 561, L31 

Fang, T., Bryan, G.L. & Canizares, C.R. 2002, ApJ, 564, 604 

Fang, T. & Canizares, C. R. 2000, ApJ, 539, 532 

Fang, T. et al. 2002, ApJ, 572, L127 

Ferland, G.J. et al. 1998, PASP, 110, 761 

Fukugita, M., Hogan, C. J., & Peebles, P. J. E. 1998, ApJ, 503, 518. 

Houck, J. C. & Denicola, L. A. 2000, ASP Conf. Ser. 216: Astronomical Data Analysis 

Software and Systems IX, 9, 591 
Heckman, T. M. et al 2002, ApJ, submitted (astro-ph/0205556) 
Hellsten, U., Gnedin, N. Y., & Miralda-Escude, J. 1998, ApJ, 509, 56 
Nicastro, F. et al. 2002, ApJ, 573, 157 

Penton, S. V., Shull, J. M., & Stocke, J. T. 2000, ApJ, 544, 150. 

Perna, P. & Loeb, A. 1998, ApJ,503, L135 

Rasmussen, J. & Pedersen, K. 2001, ApJ, 559, 892 

Rasmussen, A. et al. 2002, this proceeding 

Sembach, K. R. et al. 2001, ApJ, 561, 573 

Shapiro, P. R. & Bahcall, J. N. 1981, ApJ, 245, 335 

Shull, J. M. et al. 1998, AJ, 116, 2094 

Sutherland, R. S. & Dopita, M. A. 1993, ApJS, 88, 253 

Tripp, T.M. & Savage, B.D. 2000, ApJ, 542, 42